|Publication number||US7881932 B2|
|Application number||US 11/537,769|
|Publication date||Feb 1, 2011|
|Filing date||Oct 2, 2006|
|Priority date||Oct 2, 2006|
|Also published as||US20080082963|
|Publication number||11537769, 537769, US 7881932 B2, US 7881932B2, US-B2-7881932, US7881932 B2, US7881932B2|
|Inventors||Brien H. Muschett|
|Original Assignee||Nuance Communications, Inc.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (30), Non-Patent Citations (1), Referenced by (3), Classifications (11), Legal Events (3)|
|External Links: USPTO, USPTO Assignment, Espacenet|
1. Field of the Invention
The present invention relates the VoiceXML software language and, more particularly, a VoiceXML language extension for natively supporting voice enrolled grammars.
2. Description of the Related Art
Voice enrolled grammars are an industry standard speech recognition feature that permits users to enroll phrases into a recognition grammar using their voice. For example, a list of contacts can be created and maintained by recording spoken contact names. The spoken names are added to a speaker dependent recognition grammar, which can thereafter recognize future utterances of the names. This technique is sometimes referred to as speaker-dependent recognition. Use of voice enrolled grammars permits users to vocally customize personal services, such as Voice Activated Dialing services.
Voice enrollment often utilizes enrollment sessions to add, remove, or update acoustically provided phrases to an associated speech grammar, called a personal grammar. In an enrollment session, new acoustically provided phrases can be added to a personal grammar by initially providing a phrase, which may need to be repeated numerous times to ensure consistency is achieved for the phrase. If the new phrase does not clash with existing phrases in the personal grammar, a new phrase entry can be generated. The phrase entry can thereafter be speech recognized.
A voice enrolled grammar, unlike a text-based grammar, can be used in situations where a textual representation of a phrase is unavailable. This is often the case when a phrase represents a proper noun, such as a name. The following dialog example illustrates such a situation.
Add a person to my address book
Please say the name you would like to add or say cancel.
What is the phone number for [Bob Smith]?
Five five five one two one two
*return to top level menu*
Call Bob Smith
Calling [Bob Smith]
*System dials 555-1212*
In the above example, voice enrollment can be used to add an acoustically provided phrase of [Bob Smith] to a personal grammar. It should be noted that the acoustically provided phrase is used not only to enroll [Bob Smith] but is recorded and presented back to the user in other interactions.
Conventional techniques for implementing voice enrolled grammars and related technologies in VoiceXML have significant shortcomings. These shortcoming result from a lack of native support for voice enrolled grammars in the VoiceXML language model. Existing solutions implementing voice enrolled grammars with VoiceXML based products use the Web programming model and/or a VoiceXML object tag to access voice enrollment features, neither of which provide native language access. This results in many programming inefficiencies and a lack of uniform handling of speech recognition tasks when using both text enrolled grammars and voice enrolled grammars.
This present invention extends the VoiceXML language model to natively support voice enrolled grammars. Specifically, three VoiceXML tags can be added to the language model to add, modify, and delete acoustically provided phrases to voice enrolled grammars. Once created, the voice enrolled grammars can be used in normal speaker-dependent speech recognition operations. That is, the voice enrolled grammars can be referenced and utilized just like text enrolled grammars can be referenced and utilized. For example, using the present invention, voice enrolled grammars can be reference by standard text-based Speech Recognition Grammar specification (SRGS) grammars to create more complex, usable grammars.
The present invention can be implemented in accordance with numerous aspects consistent with material presented herein. For example, one aspect of the present invention can include a VoiceXML language that includes voice enrolled grammar language elements as part of its language mode. The elements can include an addvoicephrase, a deletevoicephrase, and/or an updatevoicephrase element. The addvoicephrase element can be a VoiceXML element configured to add a voicephrase to a voice enrolled grammar. The deletevoicephrase element can be a VoiceXML element configured to delete a voice phrase from an existing voice enrolled grammar. The updatevoicephrase element can be a VoiceXML element configured to update a voice phrase included in an existing voice enrolled grammar. The language elements can permit voice enrolled grammars to be referenced by standard text-based grammars, such as a SRGS compliant grammar.
Another aspect of the present invention can include VoiceXML code for manipulating voice phrases contained within a voice enrolled grammar. The VoiceXML code can be stored on a medium accessible by a computing device, which is capable of executing programmatic instructions of the VoiceXML code. These programmatic instructions can cause the computing device to perform one or more programmatic actions relating to a voice enrolled grammar.
The VoiceXML code can include a VoiceXML native tag selected from a list of tags consisting of an addvoicephrase tag, a deletevoicephrase tag, and an updatevoicephrase tag. The code can also include an identifier associated with the VoiceXML native tag that identifies a voice phrase. Execution of the VoiceXML code can result in a change to the voice enrolled grammar relating to the identified voice phrase. Specifics of the change are determined be specifics associated with the VoiceXML native tag. The addvoicephrase tag can cause the identified voice phrase to be added to the grammar. The deletevoicephrase tag can cause the identified voice phrase to be deleted from the grammar. The updatevoicephrase tag can cause a grammar entry for the identified voice phrase to be updated with a new voice phrase.
It should be noted that various aspects of the invention can be implemented as a program for controlling computing equipment to implement the functions described herein, or a program for enabling computing equipment to perform processes corresponding to the steps disclosed herein. This program may be provided by storing the program in a magnetic disk, an optical disk, a semiconductor memory, or any other recording medium. The program can also be provided as a digitally encoded signal conveyed via a carrier wave. The described program can be a single program or can be implemented as multiple subprograms, each of which interact within a single computing device or interact in a distributed fashion across a network space.
There are shown in the drawings, embodiments which are presently preferred, it being understood, however, that the invention is not limited to the precise arrangements and instrumentalities shown.
More specifically, during enrollment session 150, acoustically provided voice phrase (Phrase_A) can be provided to enrollment server 120 along with a request to add the phrase to a voice enrolled grammar 130. The voice enrolled grammar 130 can be a personal grammar or a speaker dependent grammar associated with speaker 105. Grammar 130 can originally include Phrase_B, Phrase_C, Phrase_D. VoiceXML code 122, which can include a VoiceXML native tag 123 for adding a voice phrase, can cause Phrase_A to be added to the grammar 130, which results in modified grammar 132. Grammar 132 can include Phrase_A, Phrase_B, Phrase_C, and Phrase_D.
During a different enrollment session 152, an acoustically provided phrase (Phrase_D) can be provided to enrollment server 120 along with a request to delete the identified phrase from grammar 132. VoiceXML code 124, which can include a VoiceXML native tag 125 for deleting a voice phrase, can cause Phrase_D to be removed from the grammar 132, which results in modified grammar 134. Grammar 134 can, thus, include Phrase_A, Phrase_B, Phrase_C.
Finally, during enrollment session 154, an acoustically provided phrase (Phrase_B′) can be provided to enrollment server 120 along with a request to update Phrase_B using Phrase_B′. VoiceXML code 126, which can include a VoiceXML native tag 127 for updating a voice phrase, can cause Phrase_B to be modified, which results in modified grammar 136. Grammar 136 can, thus, include Phrase_A, Phrase_B′, and Phrase_C.
It should be noted that the audio source 110 can include a set of previously recorded audio phrases, such as dialog phrases previously uttered by speaker 105 during one or more dialog sessions. In one configuration of system 100, modifications to the grammars 130-136 can occur during a dialog session based upon input directly provided by speaker 105. In another configuration, modifications to the grammars 130-136 can be handled as batch processes, which receive input from audio source 110. This input can include phrases previously uttered by speaker 105.
As previously noted, the speech recognition engine 140 can include a date store 146 that contains the voice enrolled grammar 136. Engine 140 can include an acoustic model processor 142 and a language model processor 144. The language model processor 144 can utilize words, phrases, weights, and rules defined by grammars 136 of data store 146. The language processor 144 can be configured to handle grammar-based language model grammars as well as statistical language model grammars.
Once enrolled, the voice enrolled grammar 130-136 can be referenced and used by the speech recognition engine 140 just like a text-based grammar. Accordingly, voice enrolled grammars 136 can be referenced by any standard text-based grammar, such as a Speech Recognition Grammar Specification (SRGS) compliant grammar to create more complex, usable grammars.
It should be realized that components of system 100 can be communicatively linked via one or more networks (not shown). The networks can include any hardware/software/and firmware necessary to convey digital content encoded within carrier waves. Content can be contained within analog or digital signals and conveyed through data or voice channels. The networks can include local components and data pathways necessary for communications to be exchanged among computing device components and between integrated device components and peripheral devices. The networks can also include network equipment, such as routers, data lines, hubs, and intermediary servers which together form a packet-based network, such as the Internet or an intranet. The networks can further include circuit-based communication components and mobile communication components, such as telephony switches, modems cellular communication towers, and the like. The networks can include line based and/or wireless communication pathways.
Each of the elements 212-216 in table 200 are elements 210 of a VoiceXML language mode. The elements 210 can include an addvoicephrase 212 element, a deletevoicephrase 214 element, and an updatevoicephrase 216 element. Each element 212-216 can have a defined set of attributes 220, returns 230, and associated events 240. All attribute 220 values can correspond to VoiceXML naming conventions. Additionally, phrase related values used for speech recognition grammars can conform to the naming rules specified for by a Speech Recognition Grammar Specification (SRGS) compliant grammar. The invention is not limited in this regard, and the phrases used for speech recognition can conform to other similar specifications, such as a NUANCE Grammar Specification Language (GSL) specification, a JAVA Speech Grammar Format (JSGF) specification, and the like, depending upon implementation specifics.
The addvoicephrase 212 element can be a VoiceXML input item that can be used to add a phrase to a new or existing voice enrolled grammar. The addvoicephrase 212 element can collect voice input and can generate a phrase entry in the personal grammar if the input does not clash with existing phrases or with entries associated with a set of bed phrases associated with a badphrasesuri attribute. In many cases, input can be collected multiple times until either it's consistent or until the input has exceeded a maximum number of attempts, which can be indicated by a maxattempts attribute.
The addvoicephrase 212 input item can provide two methods to obtain input. The first method, which can be a system default, can use normal input collection in the caller's voice can be used for phrase enrollment. The second method can use a set of acoustic phrases associated with an audioinputurilist attribute. This attribute can specify one or more URI's for which audio can be retrieved. Retrieved audio can then be used for phrase enrollment purposes. This includes audio references resulting from use of the VoiceXML record tag or the VoiceXML recordutterance property.
The addvoicephrase 212 can collect input from an input source referenced by the audioinputurilist attribute for each utterance collection necessary to achieve consistency. If the list length cannot satisfy the consistency requirement then an error event associated with the maximum number of attempts (e.g. error.enroll.maxattampts) can be triggered. This triggered event can be generated unless the minconsistent attribute is set to a value of one. Prompting can continuously occur up to number of attempts reference by the maxattempts attribute so the audio source referenced by the audioinputurilist attribute can be used each time the prompt count is increased.
It should be noted that the audio source referenced by the audioinputurilist attribute can be used for batch mode operation where audio has previously been collected. Hence, enrollment can be automated. In such an arrangement, the maxattempts attribute and the minconsistent attribute can be set to a value of one and attributes for clash and consistency levels can be set to the most lenient values. The audioinputurilist attribute can reference one audio source. If the application leave the enrollment input item, enrollment can be aborted and the enrollment grammar can remain unchanged.
It should also be noted that the addvoicephrase 212 element can support the same sub-elements as the other VoiceXML 2.0 input items with the exception of those elements related to speech and DTMF input (e.g., grammar, link, option). Local grammars may not be permitted. Global grammars can be used to terminate phrase enrollment and in some cases may result in an assignment to the input item name. In addition, the behavior of an input collection phase can be identical to the VoiceXML input collection phases as defined in Appendix C of the VoiceXML 2.0 specification. In addition to the attributes 222, the addvoicephrase 212 element can support additional VoiceXML attributes, such as cond, modal, and expr, as defined for input items.
The attributes 222 associated with the addvoicephrase 212 element can include attributes of name, phraseid, phraseidexpr, src, srcexpr, phrasenl, phrasenlexpr, weight, recordphraseaudio, consistencylevel, classlevel, maxattempts, minconsisistent, badphrasesuri, audioinputurilist, and other such attributes. The name attribute can be a form item variable in the dialog scope that holds a result. The name attribute should be unique among form items in the form. If the name is not unique, then a badfetch error can be thrown when a related document is fetched. Upon successful enrollment, a value of the name attribute can be the phraseid for an enrolled voice phrase. Shadow variables can be assigned depending on a value of the recordphraseaudio attribute.
The phraseid attribute can be a required attribute that represents an id of a voice phrase being added. An alternative to the phraseid attribute is the phraseidexpr attribute, which is associated with a phrase expression instead of a reference to an expression. A value for the phraseid or phraseidexpr can be returned in application.lastresult$.utterance when a personal grammar is used for normal speech recognition. If neither the phraseid nor phraseidexpr is specified, a badfetch error and be thrown.
The src attribute can be a required attribute that represents the new or existing personal grammar that will be updated with the new phraseid. An alternative to the src attribute is the srcexpr attribute, which is associated with a grammar's identifier instead of a reference to a grammar identifier. The grammar can be updated at the URI using an HTTP PUT request. If neither the src nor the srcexpr is specified, a badfetch error can be thrown.
The phrasenl attribute can be an optional attribute used to specify a semantic tag value to associate with the phraseid. Phrasenlexpr can be used in place of phrasenl. A value for the phrasenl or phrasenlexpr attribute can be returned in application.lastresult$.interpretation when a personal grammar is recognized. If neither the phrasenl nor the phrasenlexpr attribute is specified, a badfetch error is thrown.
The weight attribute can be an optional attribute that specifies a phrase weight. The weight attribute can be implemented as a float type having a value form 0.0 to 1.0.
The recordphraseaudio attribute can be an optional attribute that is used to specify whether an associated phrase recording is returned upon successful enrollment. Values for the recordphraseaudio attribute can include a true and a false value, where false can be a default. When a value of recordphraseaudio is true, then shadow variables for bestrecording, bestrecordingsize, and bestrecordingduration can be assigned to the input item variable upon successful enrollment. The bestrecording variable can store a reference to the recording, or remain undefined if no audio is collected. Implementation for the bestrecording variable can be platform specific. The bestrecordingsize variable can store a size of the recording (in bytes), or can remain undefined if no audio is collected. The bestrecordingduration variable can store a duration of the recording (in milliseconds), or undefined if no audio is collected. When a value of recordphraseaudio is false, then the shadow variables can be undefined and an input item variable can be assigned a value of the phraseid that was enrolled.
The consistencylevel attribute can be an optional attribute used to specify a level of similarity needed between utterances to obtain a consistent pronunciation.
A value of the consistencylevel attribute can be between 0.0 and 1.0. The higher the level of the consistencylevel attribute, the closer a match. The consistencylevel attribute can defaults to a platform specific value.
The clashlevel attribute can be an optional attribute used to specify a level of similarity between two different pronunciations that is needed to be unambiguous.
A value of the clashlevel attribute can be between 0.0 and 1.0. The higher value for the clashlevel attribute can represent an increased opportunity for clashes while reducing a likelihood of recognition ambiguity among other phrases. A lower value for the clashlevel attribute can reduce clashes but can increase a likelihood of recognition ambiguity among other phrases. The clashlevel attribute can default to a platform specific value. A value of ‘0.0’ for the clashlevel attribute can signify that a clash test will not be performed.
The maxattempts attribute can be an optional attribute used to specify the maximum number of training attempts that will be used to achieve a consistent statistical model. When a value of the maxattempts attribute is exceeded an “error.enroll.maxattempts” event can result and the input item variable can be assigned the number of attempts performed. A default value for the maxattempts attribute can be four (4). A range of values for the maxattempts attribute can be 0<minconsistent<=maxattempts or an error badfetch can be thrown.
The minconsistent attribute can be an optional attribute used to specify the minimum number of consistent pronunciations that must be obtained to voice enroll a phrase. A default value for the minconsistent attribute can be two (2). A range of values for the minconsistent attribute can be one (1) to maxattempts.
The badphrasesuri attribute can be an optional attribute used to specify an SRGS grammar that contains phrases that should not be spoken as part of the enrollment. Phrases that match this grammar can generate a clash event.
The audioinputurilist attribute can be an optional attribute used to specify the location of audio to use for batch mode phrase enrollment of enrollment where the audio is collected in a separate dialog using VoiceXML record or ‘recordutterance’. The attribute type can be a comma separated URI list of one (1) or more elements used to satisfy the collection of a phrase up to minconsistent or maxattempts. When the URI list is exhausted before consistency is reached, a maxattempts event can result. To batch enroll, maxattempts should set to one (1), minconsistent should be set to one (1) and one audio URI should be exist.
The addvoicephrase 212 element can return one or more object 232 that is either a complex or simple result depending upon an enrollment outcome. The return object 222 can include a phraseid object, a maxattempts object, an audio.ref object, an audio.duration object and an audio.size object. The phraseid object can specify a successfully enrolled phraseid, when the recordphraseaudio attribute is false. The maxattempts object can return a number of attempts made whenever an enrollment session fails due to an occurrence of a maximum number of attempts without achieving a minimum consistency among voice phrases. The audio.ref object can include an audio URI when the recordphraseaudio attribute is true. The audio.ref object can include a duration of audio referenced by the audio.ref object. The duration can be in milliseconds by default. The audio.size object can include a length a audio, which can be in bytes by default.
The addvoicephrase 212 element can situationally generate events 242 of error.enroll.clash and error.enroll.maxattempts. The error.enroll.clash event can be generated when a clash is detected between the voice phrase and an existing phrase in the voice enrolled grammar (reference by grammar URI) or between the voice phrase and a phrase contained in a set of bad grammar phrases (referenced by the badphrasesuri attribute). The value of the message associated with this event, which can be returned in object 232, can be a comma delimited list of phraseids and a comma delimited list of confusable phrases. The error.enroll.maxattempts event can be generated when consistency is not achieved within a maximum number of attempts, as determined by the maxattempts attribute. The value of a message associated with this event, which can be returned in object 232, can be the maxattempts count.
The deletevoicephrase 214 element can be a VoiceXML executable item that is used to delete a phrase from an existing voice enrolled grammar. The deletevoicephrase 214 element can be associated with numerous attributes 224, which include a phraseid attribute, phraseidexpr attribute, a src attribute, a srcexpr attribute, and the like.
The phraseid attribute can be a required attribute that represents the id of the phrase to be deleted. Phraseidexpr can be used in place of phraseid. The phraseid or phraseidexpr attribute can be returned in application.lastresult$.utterance when a personal grammar is used for normal speech recognition. If neither the phraseid or phraseidexpr is specified, a badfetch error can be thrown.
The src attribute can be a required attribute that represents the existing personal grammar that will be updated by removing the phraseid. Srcexpr can be used in place of src. The grammar can be updated at the URI using HTTP PUT request. When the deleted phrase is the last phrase in the grammar, an HTTP DELETE request can be sent to the server. If neither the src or srcexpr is specified, a badfetch error can be thrown.
No return object 234 is associated with the deletevoicephrase 214 element, which can be an executable item that does not return anything. The deletevoicephrase 214 element can generate an event 244 of error.enroll.deletevoicephrase when a voice phrase deletion attempt was made using the deletevoicephrase 214 element and when the phrase was unable to be deleted.
The updatevoicephrase 216 element can be an executable item used to update a phrase in an existing voice enrolled grammar. The updatevoicephrase 216 element can be associated with numerous attributes 226, which include a phraseid attribute, phraseidexpr attribute, a newphraseid attribute, a newphraseidexpr attribute, a src attribute, a srcexpr attribute, a phrasenl, a phrasenlexpr attribute, a weight, and the like.
The phraseid attribute can be a required attribute that represents the id of the phrase to be updated. Phraseidexpr can be used in place of phraseid. The phraseid or phraseidexpr attribute can be returned in application.lastresult$.utterance when a personal grammar is used for normal speech recognition. If neither the phraseid nor phraseidexpr is specified, a badfetch error can be thrown.
The newphraseid attribute can be a required attribute that represents the id of the new phrase to be added. Newphraseidexpr can be used in place of newphraseid. The newphraseid or newphraseidexpr attribute can be returned in application.lastresult$.utterance when a personal grammar is used for normal speech recognition. If neither the newphraseid nor the newphraseidexpr is specified, a badfetch error can be thrown.
The src attribute can be a required attribute that represents an existing personal grammar that will be updated with the new phraseid. Srcexpr can be used in place of src. The grammar can be updated at the URI using HTTP PUT request. One of src or srcexpr must be specified or a badfetch error is thrown.
The phrasenl can be an optional attribute used to specify a semantic tag value to associate with the phrase ID. Phrasenlexpr can be used in place of phrasenl. The phrasenl or phrasenlexpr attribute can be returned in application.lastresult$.interpretation when a personal grammar is recognized. One of phrasenl or phrasenlexpr must be specified or a badfetch error is thrown.
The weight attribute can be an optional attribute used to specify the phrase weight as a float from 0.0 to 1.0.
No return object 236 is associated with the updatevoicephrase 216 element, which can be an executable item that does not return anything. The updatevoicephrase 216 element can generate an event 246 of error.enroll.updatevoicephrase when a voice phrase update attempt was made using the updatevoicephrase 216 element and when the phrase was unable to be updated.
The present invention may be realized hardware, software, or a combination of hardware and software. The present invention may be realized in a centralized fashion in one computer system or in a distributed fashion where different elements are spread across several interconnected computer systems. Any kind of computer system or other apparatus adapted for carrying out the methods described herein is suited. A typical combination of hardware and software may be a general purpose computer system with a computer program that, when being loaded and executed, controls the computer system such that it carries out the methods described herein.
The present invention also may be embedded in a computer program product, which comprises all the features enabling the implementation of the methods described herein, and which when loaded in a computer system is able to carry out these methods. Computer program in the present context means any expression, in any language, code or notation, of a set of instructions intended to cause a system having an information processing capability to perform a particular function either directly or after either or both of the following: a) conversion to another language, code or notation; b) reproduction in a different material form.
This invention may be embodied in other forms without departing from the spirit or essential attributes thereof. Accordingly, reference should be made to the following claims, rather than to the foregoing specification, as indicating the scope of the invention.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5917891 *||Oct 7, 1996||Jun 29, 1999||Northern Telecom, Limited||Voice-dialing system using adaptive model of calling behavior|
|US6389395||Nov 1, 1995||May 14, 2002||British Telecommunications Public Limited Company||System and method for generating a phonetic baseform for a word and using the generated baseform for speech recognition|
|US6539359 *||Aug 23, 2000||Mar 25, 2003||Motorola, Inc.||Markup language for interactive services and methods thereof|
|US6708153 *||Jul 2, 2002||Mar 16, 2004||Hewlett-Packard Development Company, L.P.||Voice site personality setting|
|US6961776 *||Dec 22, 2000||Nov 1, 2005||Nortel Networks Limited||Architecture for multiple channel access to applications|
|US6985865 *||Sep 26, 2001||Jan 10, 2006||Sprint Spectrum L.P.||Method and system for enhanced response to voice commands in a voice command platform|
|US7013276||Oct 5, 2001||Mar 14, 2006||Comverse, Inc.||Method of assessing degree of acoustic confusability, and system therefor|
|US7117159 *||Sep 26, 2001||Oct 3, 2006||Sprint Spectrum L.P.||Method and system for dynamic control over modes of operation of voice-processing in a voice command platform|
|US7140004 *||Jun 13, 2000||Nov 21, 2006||Tellme Networks, Inc.||Method and apparatus for zero-footprint phone application development|
|US7180986 *||Jun 29, 2005||Feb 20, 2007||Glenayre Electronics, Inc.||Enhanced voicemail system|
|US7231636 *||Dec 21, 2001||Jun 12, 2007||Nortel Networks Limited||System and method for tracking VoiceXML document execution in real-time|
|US7313527 *||Jan 23, 2003||Dec 25, 2007||Intel Corporation||Registering an utterance and an associated destination anchor with a speech recognition engine|
|US7349836 *||Dec 12, 2003||Mar 25, 2008||International Business Machines Corporation||Method and process to generate real time input/output in a voice XML run-time simulation environment|
|US7412391 *||Nov 22, 2005||Aug 12, 2008||Canon Kabushiki Kaisha||User interface design apparatus and method|
|US7487085 *||Aug 24, 2004||Feb 3, 2009||International Business Machines Corporation||Method and system of building a grammar rule with baseforms generated dynamically from user utterances|
|US7546527 *||Mar 6, 2001||Jun 9, 2009||International Business Machines Corporation||Method and apparatus for repurposing formatted content|
|US7657828 *||Jun 5, 2006||Feb 2, 2010||Nuance Communications, Inc.||Method and apparatus for coupling a visual browser to a voice browser|
|US7739589 *||Nov 2, 2005||Jun 15, 2010||International Business Machines Corporation||Extending markup of a browser using a plug-in framework|
|US20020169604||Mar 9, 2001||Nov 14, 2002||Damiba Bertrand A.||System, method and computer program product for genre-based grammars and acoustic models in a speech recognition framework|
|US20020173964||Mar 30, 2001||Nov 21, 2002||International Business Machines Corporation||Speech driven data selection in a voice-enabled program|
|US20030009331||Jul 16, 2001||Jan 9, 2003||Johan Schalkwyk||Grammars for speech recognition|
|US20030023440||Mar 9, 2001||Jan 30, 2003||Chu Wesley A.||System, Method and computer program product for presenting large lists over a voice user interface utilizing dynamic segmentation and drill down selection|
|US20030139925 *||Dec 31, 2001||Jul 24, 2003||Intel Corporation||Automating tuning of speech recognition systems|
|US20040194016 *||Mar 28, 2003||Sep 30, 2004||International Business Machines Corporation||Dynamic data migration for structured markup language schema changes|
|US20040199375||Apr 5, 2004||Oct 7, 2004||Farzad Ehsani||Phrase-based dialogue modeling with particular application to creating a recognition grammar for a voice-controlled user interface|
|US20050171775||Mar 29, 2005||Aug 4, 2005||Sean Doyle||Automatically improving a voice recognition system|
|US20060047510||Aug 24, 2004||Mar 2, 2006||International Business Machines Corporation||Method and system of building a grammar rule with baseforms generated dynamically from user utterances|
|US20060074652||Aug 9, 2005||Apr 6, 2006||International Business Machines Corporation||Method and system for voice-enabled autofill|
|US20060136222||Dec 22, 2004||Jun 22, 2006||New Orchard Road||Enabling voice selection of user preferences|
|US20060230410 *||Mar 22, 2006||Oct 12, 2006||Alex Kurganov||Methods and systems for developing and testing speech applications|
|1||"Automatic Speech Recognition", Loquendo, AMBR 042 ENG, Jan. 2006.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US9148499||Jan 22, 2013||Sep 29, 2015||Blackberry Limited||Method and system for automatically identifying voice tags through user operation|
|US9171069 *||Jul 31, 2012||Oct 27, 2015||Freedom Solutions Group, Llc||Method and apparatus for analyzing a document|
|US20140040270 *||Jul 31, 2012||Feb 6, 2014||Freedom Solutions Group, LLC, d/b/a Microsystems||Method and apparatus for analyzing a document|
|U.S. Classification||704/244, 715/234, 704/270.1|
|International Classification||G10L15/22, G06F17/24, G10L15/06|
|Cooperative Classification||G10L2015/0631, G06F17/2247, G10L15/22|
|European Classification||G06F17/22M, G10L15/22|
|Oct 2, 2006||AS||Assignment|
Owner name: INTERNATIONAL BUSINESS MACHINES CORPORATION, NEW Y
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:MUSCHETT, BRIEN H.;REEL/FRAME:018333/0217
Effective date: 20061002
|May 13, 2009||AS||Assignment|
Owner name: NUANCE COMMUNICATIONS, INC., MASSACHUSETTS
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:INTERNATIONAL BUSINESS MACHINES CORPORATION;REEL/FRAME:022689/0317
Effective date: 20090331
Owner name: NUANCE COMMUNICATIONS, INC.,MASSACHUSETTS
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:INTERNATIONAL BUSINESS MACHINES CORPORATION;REEL/FRAME:022689/0317
Effective date: 20090331
|Jul 2, 2014||FPAY||Fee payment|
Year of fee payment: 4